How to Build Effective Data Metrics and Indicator Systems for Business Decisions

Data are raw, unprocessed records; their essence lies in using mathematics to observe, record, and understand the world, turning qualitative observations into quantitative insights.

Metrics combine data with business scenarios to guide actions. A good metric should answer the five W questions (who, when, where, what, why) and address dimensions, clear definitions, and logical relationships.

1. Usage Scenarios (Who, When, Where)

Define dimensions (e.g., gender, age, region) to clarify the analysis scenarios a metric supports.

2. Metric Definition (What)

Resolve calculation scope and avoid ambiguous naming (e.g., same name, different meaning).

3. Metric Purpose (Why)

Clarify logical relationships between metrics, such as profit = revenue – cost – taxes – refunds.

4. Building an Indicator System

Two widely used frameworks are introduced: Pirate Metrics (AARRR) and the First Key Metric method (also known as the North Star metric).

Pirate Metrics (AARRR)

Proposed by Dave McClure in 2007, it splits metrics into Acquisition, Activation, Retention, Revenue, and Referral, forming a chain that guides growth‑focused analysis.

First Key Metric Method

The core idea is that at any moment there is one most critical metric, which may change as the business evolves through stages such as MVP, growth, and revenue, and across models like e‑commerce, SaaS, mobile apps, or marketplaces.

Steps to construct the system (illustrated with a case study):

1. Identify the First Key Metric

For a cross‑border e‑commerce leader, revenue (sales) was chosen despite being a “vanity” metric, because it drives overall business goals.

2. Divide into Modules

Beyond user conversion and retention, modules cover procurement, warehousing, logistics, cost, and timeliness.

3. Map Logical Relationships

Starting from the first key metric, map how each sub‑metric contributes to it.

Different industries and development stages will produce varied “metric trees.” Choose the framework that fits the context, but always ensure metrics provide clear decision‑making support and unlock data value.